I understand how frustrating it can be when the TTS merges words together. A great way to tackle this is by customizing the pronunciation using some pre-processing techniques before sending the text to the TTS engine. Here are a few friendly options you might find helpful: • Using the

before_tts_cb

Callback: You can tweak the text just before it reaches the TTS model. For instance, replacing tricky words with a phonetically clearer version often does the trick. This can be done smoothly in Python or Node.js. • Using SSML (Speech Synthesis Markup Language): This lets you set custom rules for pronunciation, pitch, and speed. Most TTS providers support this, giving you a lot of flexibility. • Provider-Specific Customizations: Some TTS plugins, like Google Cloud TTS, even let you define custom pronunciations directly through their API, which can be really handy. If you’d like, I’m here to help with specific examples or walk you through implementing any of these solutions. Just let me know! Sources: VoicePipelineAgent | LiveKit Docs | Agents v0.x migration guide - Node.js | LiveKit Docs | Agent speech and audio | LiveKit Docs | livekit-plugins/livekit-plugins-cartesia/livekit/plugins/cartesia/tts.py